AI Glossary

systems

Agent

AI Agent

An AI that can plan and take actions, not just answer questions

Reading level

PRACTITIONER — Technical context

AI agents combine an LLM (for reasoning and planning) with tool use and a feedback loop. The agent observes its environment, reasons about what actions to take, executes tool calls (web search, code execution, API calls), observes the results, and iterates until the goal is achieved. Common frameworks: ReAct (reason+act interleaved), Plan-and-Execute, multi-agent systems with specialized sub-agents.

Real-world example

Devin (the AI software engineer) is an agent. You give it a GitHub issue, and it reads your codebase, writes code, runs tests, fixes failures, and opens a pull request — all autonomously.

systemsagentsagentic AI